Spearman correlation

Spearman Rank Correlation

A rank-based correlation coefficient — converts values to ranks and measures correlation between ranks rather than raw values, making it robust to outliers and non-linear monotonic relationships.

When you'd see it: When variables are ordinal (satisfaction ratings, ranking data) or when the relationship between continuous variables is monotonic but not linear. Commonly used in customer satisfaction analysis, academic ranking studies, and any analysis where exact values are less reliable than rank order.

Why it matters: Spearman makes fewer assumptions than Pearson — it doesn't require linear relationships or normally distributed data. For business data that's skewed or ordinal, Spearman is usually more appropriate than Pearson, even though Pearson is the default in most textbooks.

Common mistakes: Using Pearson correlation when the data is ordinal. A customer giving a score of 7 vs 8 on a 10-point scale doesn't imply the same interval as a score of 2 vs 3. Pearson treats the differences as equal; Spearman only cares about relative rank, which is usually the right assumption for survey data.

Related Pearson correlation, correlation, covariance